Amidst a devastating civil war, the food crisis in Sudan is deteriorating rapidly.

According to the World Food Programme (WFP), 18 million people are enduring acute food insecurity, with nearly 5 million suffering at emergency levels—the highest number ever recorded during the harvest season. The Nutrition Cluster in Sudan has assessed that 729,000 children under five are suffering from severe acute malnutrition, the deadliest form of extreme hunger. Clingandael reports that deaths from malnutrition or starvation have been recorded in central Darfur and parts of Khartoum.

An even deeper catastrophe looms. As Anette Hoffman writes in the Clingandael policy brief, “this is usually the season when food is more available and affordable than during the remainder of the year. The current food crisis is therefore expected to become much worse in the coming months.” Save the Children has warned that nearly “230,000 children, pregnant women and new mothers could die” in that period. According to Clingandael’s most optimistic assessment, most of the population will face extreme hunger by July, with 6 percent facing catastrophic hunger, including possible famine conditions in Khartoum and Darfur. In more pessimistic scenarios, there may be famine in much of the country or even across its entirety. In sum, “Sudan’s trajectory towards a catastrophe is certain; the risk of famine is very high.” The WFP has warned that the situation risks becoming “the world’s worst hunger crisis.” The United Nations (UN) Secretary-General is poised to alert the Security Council to the gravity of the situation, pursuant to his responsibility under paragraph 12 of Security Council Resolution 2417, the Council’s framing resolution for oversight and enforcement regarding starvation in armed conflict.

There is no doubt that the civil war between the Sudanese Armed Forces (SAF), led by General Abdel Fattah al-Burhan, and the Rapid Support Forces (RSF), led by General Mohamed Hamdan ‘Hemedti’ Dagalo, is a key driver of the crisis. But hunger is not simply a collateral consequence of the conflict. It is the product of those belligerents’ choices about how to conduct it.

“Food and hunger have undeniably become weapons,” with both the SAF and the RSF “destroying and looting food infrastructure and preventing people from accessing food.” Deprivation is also being inflicted through the deliberate impediment of humanitarian access. A WFP spokesperson explained, “[W]e are not able to deliver assistance across conflict lines. … We are not getting security guarantees and the approvals to move commodities freely.”

The strategy is not covert. In February, General al-Burhan promised to cut off RSF-controlled areas completely, explaining that humanitarian access:

will not happen until we end this war and defeat these criminal rebels and drive them from the homes of citizens, government and service institutions, and from all the cities they have plundered—Nyala, El Geneina, Madani, and Khartoum.

In early March, Foreign Minister Ali El Sadig resisted calls for cross-border humanitarian aid delivery from Chad. The government ultimately bowed to intense pressure on that point. However, Médecins Sans Frontières insists that the government’s framework for access continues to “leave vast areas in Darfur, Kordofan, Khartoum and Jazeera states still inaccessible,” including due to its “systematically withhold[ing] authorisations for movements across frontlines.”

To be clear, as with other modalities of deprivation, the problem of siege starvation is not specific to the SAF—the RSF is also reported to have employed the tactic. Although focusing primarily on the policy of encirclement deprivation articulated in al-Burhan’s declaration, the analysis below applies equally to starvation crimes on both sides. It is axiomatic that the atrocities of one party have no bearing on either the rules applicable to their adversaries or the civilian status of the populations under their control.

Ultimately, it is clear that starvation war crimes are being committed in Sudan. However, for the reasons discussed below, the International Criminal Court’s (ICC’s) jurisdiction over them is not straightforward. Referrals by a suitably representative coalition of States parties could bolster the Court’s legitimacy in this context, but given enduring jurisdictional limits and uncertainties, it is also worth exploring alternative routes to accountability.

The Starvation War Crime

The starvation war crime is codified in article 8(2)(b)(xxv) of the ICC Statute for international armed conflicts (IACs) and, by amendment, article 8(2)(e)(xix) for non-international armed conflicts (NIACS), such as the ongoing conflict in Sudan (notwithstanding its transnational connections). It is also a customary war crime (see Part III). The crux of the prohibition is the intentional use of starvation of civilians as a method of warfare by depriving them of food, water, and other objects indispensable to their survival.

In clarifying what it means to starve civilians as a method of warfare, it is critical to emphasize that the presence of combatants does not change the civilian character of a population that is predominantly composed of civilians (e.g. Additional Protocol I, article 50(3); ICTY Prosecutor v. Karadžić Trial Judgment 2016, paras. 474, 4610 n.5510 ). As such, the population in each of the areas currently under RSF control is a civilian population, notwithstanding RSF presence and control. It is also important to clarify that “starvation” has transitive and intransitive meanings—referring both to the transitive act of deliberate sustenance denial and to the intransitive state of severe malnutrition at which life becomes impossible to sustain (see Part V). A final point worth clarifying is that “intent” in the ICC regime can also take alternative forms, including either seeking a particular outcome or acting in the knowledge that it will occur with a virtual certainty (ICC Statute, article 30(2)).

Taking these points together (for more detail, see here), impeding aid to RSF-controlled areas for the purpose of sustenance denial is a deliberate denial of sustenance to a civilian population. Even when undertaken with a view to squeezing embedded adversary forces, directing that kind of deprivation at a population that is properly characterized as civilian necessarily implicates the war crime of starvation of civilians as a method of warfare. Put another way, the motive for directing sustenance denial at a civilian population cannot recharacterize, and should not obscure, the intentionality of that directed sustenance denial. In the current case, available information suggests the crime is implicated on that basis.

Even if the purpose of impeding aid, obstructing coping mechanisms, or destroying food systems were something other than sustenance denial, engaging in such acts in the knowledge that civilians are virtually certain to be left in a state of starvation would itself implicate the starvation war crime. Given current levels of food insecurity, and the real prospect of famine, this too appears to be a sufficient basis for the crime to attach in Sudan.

ICC Jurisdiction

Sudan is not an ICC State Party, but the UN Security Council referred the situation in Darfur to the Court in 2005, thereby triggering the latter’s jurisdiction (via article 13(b) of the ICC Statute) and avoiding the ordinary imperative to establish that the relevant crimes either occurred on State Party territory or were perpetrated by State Party nationals.

However, absent an unlikely second and updated Security Council referral, ICC jurisdiction is not straightforward in the current context. First, the temporal scope of the Court’s jurisdiction in Security-Council-referred situations requires clarification. Second, the territorial scope of the specific referral relating to Darfur may exclude some of the starvation (and other) crimes occurring in the Sudanese conflict. Third, and most vexing, material jurisdiction over the starvation crime is complicated by the fact that the relevant Security Council referral occurred long before the amendment incorporating that crime into the ICC Statute.

The Temporal Jurisdiction of Security Council Referrals

Last July, in his periodic report to the Security Council, ICC Prosecutor Karim Khan invoked the 2005 referral to assert continuing ICC jurisdiction over international crimes committed in the region today. Specifically, he stated:

I do wish to be very clear that our mandate pursuant to Security Council Resolution 1593 is ongoing with respect of crimes within our jurisdiction: the crimes of genocide, crimes against humanity, and war crimes. …. We have already started investigating[.]

He has since declared “grounds to believe” that crimes are being committed by both the SAF and the RSF.

When the Court approved an arrest warrant in the Libyan context for conduct occurring half a decade after the relevant Security Council referral, it emphasized that the crimes were “associated with the ongoing armed conflict underlying the referral by the Security Council.” (para. 23) A similar line of analysis may prevail in the Darfur situation, where the fact of an ongoing NIAC in the region since prior to the 2005 referral (RULAC) may be deemed sufficient for continued jurisdiction. A potential complication here is that the NIAC between the previously allied SAF and RSF began only in 2023, and is a distinct dyad, demanding its own analysis (RULAC). Nonetheless, it would not be surprising if the Court were to affirm ongoing temporal jurisdiction on the grounds that the situation of insecurity that prompted the referral has not yet been resolved. Across both Security Council-referred situations, the Prosecutor has consistently maintained the enduring effect of the referrals, notwithstanding changed conditions on the ground.

The Territorial Scope of the Darfur Referral

Most ICC situations are defined with reference to a specific state. However, in this context, the Security Council referred “the situation in Darfur,” not the situation in Sudan as a whole (para. 1). Even understood in narrow terms, this would cover many of the recent and ongoing starvation crimes. For example, al-Burhan’s promise of encirclement deprivation included specific cities in Darfur (Nyala and El Geneina) in addition to much of Darfur under the general category of “RSF-controlled areas.” Darfur has also been the location of many alleged RSF starvation crimes (Clingandael Brief, p.3). However, neither party’s war crimes appear to have been limited to Darfur. Al-Burhan’s declaration denying humanitarian access included Khartoum and Madani.

On the issue of territorial jurisdiction, the Pre-Trial Chamber in Abd-Al-Rahman (‘Ali Kushayb’) authorized that case to proceed on the basis that the alleged events took “place within the perimeters of the territory of Darfur, Sudan.” (para. 8) In assessing what it means for a crime to take place in a location, it is important to note that the ICC has elsewhere declined to rule out the possibility that this may include part of the crime occurring on that territory “irrespective of whether the part occurring on the territory is a constitutive element of the crime.” (Bangladesh/Myanmar authorization, paras. 56, 62). Nonetheless, even on that approach, the Court would seem not to have jurisdiction over crimes that occur in Sudan, but entirely outside of Darfur. This would entail a significant jurisdictional gap in the current conflict.

Material Jurisdiction and the Starvation Crime

On the issue of material jurisdiction, things get even trickier. As applied to IACs, the starvation war crime has been part of the ICC’s jurisdiction from its founding. However, the crime was only incorporated for NIACs via a 2019 amendment, which entered into force in 2021. Pursuant to article 121(5) of the ICC Statute, amendments to ICC crimes operate under a distinct jurisdictional regime from the crimes as codified in the original 1998 Statute. Specifically:

Any amendment to articles 5, 6, 7, and 8 [the substantive crimes] of this Statute shall enter into force for those States Parties which have accepted the amendment one year after the deposit of their instruments of ratification or acceptance. In respect of a State Party which has not accepted the amendment, the Court shall not exercise its jurisdiction regarding a crime covered by the amendment when committed by that State Party’s nationals or on its territory.

Given this, it is not clear how to think about ICC jurisdiction vis-à-vis amendment crimes in situations in which that jurisdiction is predicated on a Security Council referral.

A reasonable argument can be made that referrals incorporate amendment crimes. A decade ago, Dapo Akande outlined the key points of such an argument as applied to a hypothetical Security Council referral of the situation in Syria and the possibility of a chemical weapons prosecution under articles 8(2)(e)(xiii) and 8(2)(e)(xiv)—themselves amendment crimes. The key elements of that argument were as follows.

First, when an amendment is adopted, it becomes part of the ICC Statute, notwithstanding the fact that its jurisdictional implications depend on specific State Party ratifications. Second, the purpose of Security Council referrals to the ICC is to expand ICC jurisdiction where its ordinary territorial or nationality jurisdiction is lacking. In other words, the point is to create jurisdiction in the absence of state consent. There is no consent-based distinction between original ICC Statute crimes and amendment crimes, so it stands to reason that a referral would bring the full Statute to bear, without differentiating between the two categories. Third, in article 15ter of the Statute, the aggression amendments provide for Security Council referral as a basis for jurisdiction over that crime. Consistency would militate in favor of the same approach to amendments to articles 6, 7, or 8. Fourth, notwithstanding the specification in article 121(5) that the Court “shall not” exercise territorial or national jurisdiction over amendment crimes where the State Party in question has not ratified the amendment, the point of that framing was to put non-ratifying States Parties on a par with states that are not party to the Statute at all. That, one might think, should have no bearing on Security Council referrals, as the latter underpin a distinct form of jurisdiction that is not predicated on direct state consent.

Akande’s analysis strikes me as likely to prevail in the kind of situation to which he was responding. The situation in Sudan, however, raises a further difficulty. How should the Court understand its material jurisdiction when that jurisdiction is predicated on a Security Council referral that predated the incorporation of the relevant amendment, or even any prospect of that amendment?

One might argue that the Council refers situations, not crimes. If the Prosecutor is correct that the referral continues to underpin ICC jurisdiction vis-à-vis crimes today and if Akande is correct that referrals include amendment crimes, why would the situation in Darfur not be governed by the ICC crimes that are applicable to Council-referred situations today? Particularly given that the amendment would apply only to crimes occurring after its entry into force (thus disposing of any objection with respect to the principle of legality) (ICC Statute, article 22(1))?

On the other hand, a standard articulation of the relationship between the ICC and the territorial state of a Security Council referral is that the referral causes the latter to function as a State Party for the purposes of the referred situation. Perhaps, per Akande’s analysis, the analogue ought to be a State Party that has ratified all components of the Statute in the moment of the referral, or at least a state that has accepted Court jurisdiction over all components of the ICC Statute (per article 12(3) of the Statute). But even if that is how one should understand what happened to Sudan in 2005 (a moment at which there were no newly incorporated crimes), a State Party from 2005 would have retained discretion over whether to ratify new amendments as and when they have been agreed. That might be thought to militate against automatically accumulating material jurisdiction in Security Council referred situations.

Two points might be raised in response to that concern. First, Resolution 1593 imposed on Sudan the duty to “cooperate fully with” the Court (para. 2). Perhaps this might be understood to include the kind of cooperation that would be equivalent to accepting the applicability of new ICC crimes as they are added to the Statute. The difficulty with this line of argument is that States Parties are all under an obligation to “cooperate fully” with the Court under article 86, and yet they clearly have freedom to decide whether to ratify amendments.

A second response might emphasize that, notwithstanding common use of the State Party analogy, Sudan is fundamentally unlike actual States Parties in several respects, one of which goes to the relevance of consent over time. Specifically, it is not just that it was brought into the Court’s ambit without consent; it also has no option to withdraw (as former States Parties, the Philippines and Burundi, have done) and thereby bring an end to the Court’s jurisdiction over its territory and nationals. With the initial referral now almost two decades old, this distinction between entry and exit is a meaningful one. If the Prosecutor is correct that the Court’s authority endures in such circumstances, it is not much of a leap to suggest that it covers the full list of statutory crimes at the moment of any particular impugned act.

Legitimacy at the ICC

At the same time, the notion of forever referrals and the prospect that they include automatically accumulating material jurisdiction is not without its costs. Indeed, even absent those exceptional features of Council-referred situations, the latter have always raised significant (and in my view well founded) legitimacy objections. The force of those objections is especially strong in the current moment, given the real possibility of the direct or indirect implication in international crimes of the three permanent Security Council members not to have ratified the ICC Statute (the members uniquely able to limit their own officials’ vulnerability to ICC authority). Moreover, the legacy on this issue is particularly raw in the Darfur context, with the Court’s arrest warrants for Omar al-Bashir the crux of its years-long standoff with the African Union (AU) and its member states.

One might object that such concerns are perfectionist and that the current imperative is for ICC action in the Sudan situation, with a view to reversing the current trajectory towards famine. The crime, after all, has customary status, so the material jurisdiction obstacle is a technicality, rather than a deep legality concern.

There is undeniable force to that objection. At the same time, the strongest evidence of the ICC’s deterrent effect indicates that it is derivative of the Court’s capacity to stigmatize, which is itself contingent on the institution’s legitimacy.

With that in mind, the Court might benefit from either a genuinely cross-regional coalition, or perhaps a robust regional coalition, of States Parties building on the practice in the Venezuela, Ukraine, and Palestine situations by supplementing the Security Council’s 2005 referral with their own current referrals under article 14 of the Statute.

Under article 14, a State Party referral would ordinarily activate the jurisdictional trigger in article 13(a), pursuant to which the Court can proceed with an investigation grounded in its ordinary jurisdiction under article 12 (i.e. on the grounds that the alleged crimes were perpetrated on the territory, or by the nationals, of a State Party or other state accepting ICC jurisdiction). In other words, unlike Security Council referrals (which activate the jurisdictional trigger in article 13(b) and are not constrained by article 12), State Party referrals cannot create jurisdiction in contexts in which the Court otherwise lacks it. In that sense, they would ordinarily be unavailable in response to situations on non-party territory involving non-party nationals, such as in Sudan. However, the only technical limit on article 14 referrals is that the situation is one in which “one or more crimes within the jurisdiction of the Court appear to have been committed.” (article 14(1)). As the Council’s referral has already created the jurisdictional basis in Darfur, that threshold is satisfied.

One might object that, for precisely that reason, State Party referrals would be legally redundant. But that is equally true of all but one referral in the each of the Venezuela, Ukraine, and Palestine situations. In those contexts, redundancy did not deter collective action. Indeed, in the latter two situations, technically redundant referrals continued to be issued even after the Prosecutor had opened the investigation. States Parties appear to have engaged in that activity in part to express solidarity with victims and bolster the Court’s legitimacy and authority in the context at hand. A similar use of expressive collective referrals by an appropriately representative coalition of states could strengthen the Court’s standing in the Sudan situation.

This is not to say such action would be politically easy. Although the AU suspended Sudan in 2021 (with ongoing effect), the legacy of the organization’s standoff with the Court over al-Bashir may preclude affirmative regional support for ICC authority in Sudan today. At the same time, the fact of that history emphasizes the heightened need for the Court to act with legitimacy in this context. Moreover, if such collective action were forthcoming, its impact would be all the greater for the fact of those past tensions.

Alternative Criminal Categories under the Rome Statute

Legitimacy aside, it may be that uncertainty regarding jurisdiction over post-referral amendments, will deter the Prosecutor from pursuing charges under the starvation war crime in Sudan. Here, the crimes against humanity of extermination and other inhumane acts, and possibly genocide, may offer an alternative jurisdictional route for prosecuting starvation crimes.

Extermination includes mass killing through the infliction of lethal conditions of life, such as through impeding aid and destroying food systems (article 7(2)(b), ICC Statute). The residual crime against humanity category of other inhumane acts attaches to operations that “intentionally caus[e] great suffering, or serious injury to body or to mental or physical health” (article 7(1)(k)). A strong case could be made for the applicability of either provision in the current context, although the latter has the advantage of avoiding the need to prove that the impugned acts caused death. To the extent it can be established that the purpose of deprivation is to destroy a racial, religious, national, or ethnic group, genocide may also apply in the form of “inflicting conditions of life calculated to bring about” that physical destruction (article 6(c)).

Alternative Jurisdictions

Finally, it may also be worth contemplating alternative criminal justice mechanisms. Most obviously, domestic courts could exercise universal jurisdiction, including over the starvation war crime. In so doing, they would be on a strong legal footing. A key motivation for the ICC amendment in 2019 was that the starvation war crime already had customary status in both IACs and NIACs (Swiss non-paper, ASP Working Group, Annex IV, paras. 3-5). Similarly, when they agreed the Malabo Protocol in 2014, African Union (AU) states included the crime for both conflict classifications (articles 28D(b)(xxvi), 28D(e)(xvi)). The Protocol recently received its first ratification.

More generally, concerns with the Security Council’s role at the ICC call for bolstering regional criminal justice efforts in response to such situations. For example, Owiso Owiso has pointed to the AU Assembly’s capacity to set up ad hoc tribunals pursuant to its authority under article 4(h) of the AU Constitutive Act—an authority the Assembly invoked in mandating Senegal to “prosecute and ensure that Hissène Habré is tried, on behalf of Africa” for crimes committed in Chad. Here, too, any such tribunal could rely securely on the customary status of starvation crimes. In a sense, this AU authority parallels that of the UN Security Council under Chapter VII of the UN Charter. However, comprising all AU member states and operating pursuant to consensus or two-thirds majority, the AU Assembly avoids some of the Security Council’s most significant legitimacy vulnerabilities.


Ultimately, the use of starvation of civilians as a method of warfare demands accountability. The question in the context of Sudan is how best to pursue it. In the meantime, the key imperative is for all of those with leverage to use it to pressure the SAF and RSF to facilitate humanitarian access, to cease using starvation as a weapon of war, and to reverse the descent into a catastrophic famine. Recognizing the criminality of starvation methods can provide the focal point for mobilizing to that end with the intensity and urgency that the situation demands.

IMAGE: Sudanese refugees and ethnic South Sudanese who have fled the war in Sudan carry their belongings while boarding a boat at the shores of the White Nile River in the Port of Renk on February 14, 2024. (Photo by LUIS TATO/AFP via Getty Images)